For context, see https://github.com/w3c/webvr/issues/173.

There's already a test to ensure that only one renderer can be presenting to a given display at a time: https://cs.chromium.org/chromium/src/device/vr/vr_display_impl_unittest.cc?sq=package:chromium&dr&l=73 

This test would likely be a derivative of that one.

I think we do still need to finish the actual implementation of this. Right now we don't listen to page focus changes to know if we're actually in the foreground. We should also be sending blur/focus events when going into the back/foreground.

mthiesse, will you please file an implementation bug?

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/eb52d480606f792b3f2c0ddfe0569eb67cd13bef

commit eb52d480606f792b3f2c0ddfe0569eb67cd13bef
Author: bsheedy <bsheedy@chromium.org>
Date: Thu Feb 23 19:31:25 2017

Add WebVR WebVR E2E test capabilities via Android instrumentation

This adds support for running WebVR tests that can't be done in layout
tests, such as end-to-end tests using the actual GVR implementation.

The tests consist of two parts: The WebVR code written in Javascript
and the test runner code written in Java.

The Javascript code uses the same testharness.js that layout tests use,
so all of its functionality such as asserts, EventWatcher, and
async_tests are available and work exactly like in layout tests.

The Java code is responsible for loading the initial HTML page and
performing actions that aren't possible through Javascript, such as
performing user gestures or opening new tabs. Additionally, the Java
code is what actually passes or fails the test based on the results
reported by testharness.js.

BUG= 679827 , 674974 

Review-Url: https://codereview.chromium.org/2588703003
Cr-Commit-Position: refs/heads/master@{#452586}

[modify] https://crrev.com/eb52d480606f792b3f2c0ddfe0569eb67cd13bef/chrome/android/BUILD.gn
[add] https://crrev.com/eb52d480606f792b3f2c0ddfe0569eb67cd13bef/chrome/android/javatests/src/org/chromium/chrome/browser/vr_shell/WebVrTest.java
[add] https://crrev.com/eb52d480606f792b3f2c0ddfe0569eb67cd13bef/chrome/test/data/android/webvr_instrumentation/html/test_nfc_fires_onvrdisplayactivate.html
[add] https://crrev.com/eb52d480606f792b3f2c0ddfe0569eb67cd13bef/chrome/test/data/android/webvr_instrumentation/html/test_pose_data_unfocused_tab.html
[add] https://crrev.com/eb52d480606f792b3f2c0ddfe0569eb67cd13bef/chrome/test/data/android/webvr_instrumentation/html/test_requestPresent_enters_vr.html
[add] https://crrev.com/eb52d480606f792b3f2c0ddfe0569eb67cd13bef/chrome/test/data/android/webvr_instrumentation/resources/webvr_boilerplate.js
[add] https://crrev.com/eb52d480606f792b3f2c0ddfe0569eb67cd13bef/chrome/test/data/android/webvr_instrumentation/resources/webvr_e2e.css
[add] https://crrev.com/eb52d480606f792b3f2c0ddfe0569eb67cd13bef/chrome/test/data/android/webvr_instrumentation/resources/webvr_e2e.js

A test for making sure getFrameData returns false and doesn't update the pose when the tab is unfocused was added in the patch that adds the WebVR E2E functionality - is this sufficient, or are there other cases we need to test?

I'm going to go ahead and close this since testPoseDataUnfocusedTab (https://cs.chromium.org/chromium/src/chrome/android/javatests/src/org/chromium/chrome/browser/vr_shell/WebVrTest.java?q=webvrtest.java+package:%5Echromium$&dr&l=307) should cover this AFAIK. If for whatever reason that's insufficient coverage, feel free to re-open or open a new bug.